Accession Number:
AD1111377
Title:
Design Theory for Chemical Detection
Descriptive Note:
[Technical Report, Memorandum Report]
Corporate Author:
NAVAL RESEARCH LAB WASHINGTON DC
Personal Author(s):
Report Date:
2020-01-01
Pagination or Media Count:
88
Abstract:
This report details a rational, quantitative approach for designing and optimizing sensor arrays for chemical detection. The work demonstrates that the degree to which sensor array capabilities and limitations can be understood is unavoidably linked to the degree to which the underlying response mechanism has been characterized, as well as inextricably linked to specific formulations of sensing task. A mathematical framework for describing chemical sensing tasks is presented and device-agnostic quality metrics based on information theoretic measures are derived and interpreted within the context of this framework. An algorithmic approach for sensor array configuration is presented, and it is demonstrated that the derived quality metrics are amenable to convex optimization. Information-theoretic approaches for quantitatively understanding the impact of uncertain chemical backgrounds on chemical detection systems, characterizing the quality of chemical simulants under realistic conditions, and understanding the underlying chemical expressiveness of a given sensor technology are derived from this framework and discussed.
Descriptors:
- computational science
- materials science
- information processing
- information science
- neural networks
- chemistry
- chemical analysis
- chemical compounds
- data mining
- detection
- bayesian networks
- chemical warfare agents
- databases
- machine learning
- materials testing
- chemical detectors
- information theory
- pattern recognition
- materials laboratories
- random variables
Subject Categories:
- Miscellaneous Detection and Detectors
- Cybernetics